Vinoba Bhave University

The National Education Policy (NEP) 2020 requires the shift to competency-based and multidisciplinary education, but due to the heterogeneity of learners and fixed pedagogical schemes, the implementation is performed on a large scale. The paper outlines a combined machine learning (ML) model of personalized instruction, which is in line with NEP 2020 goals. To achieve real-time curriculum modification, we use a hybrid method based on a supervised and unsupervised learning to multi-dimensional learner profiling, using academic, behavioral, and preferential data. Reinforcement Learning (RL) layer is adopted to optimize the instructional paths dynamically in accordance with the current interaction states and achievement results. Moreover, the framework is connected to the national digital infrastructures, such as DIKSHA, the Academic Bank of Credits (ABC) to make it scalable and multilingual. Simulated settings in terms of comparative analysis reveal that knowledge retention and learning efficiency have been greatly improved as compared to traditional static models.

The Thar Desert of northwestern India, despite its harsh ecology, has sustained settlement of ancient crafts and pastoral communities. Their persistence provides a unique opportunity to study how migration, ecology, and culture have shaped genetic diversity. We analyzed genome-wide SNP data from 176 individuals across eight occupational communities along with global Indian populations and diverse ancient genomes. Population history, ancestral migration, population structure, demography, admixture, and founder effects were elucidated using diverse population genetic statistical methods. The Thar groups occupy an intermediate position on the Indian north-south cline. Pastoralists and artisans (woodcarvers and Persian gold embossers) align with West Eurasian lineages, while potters and performers align with southern clines. Uniparental data confirmed heterogeneous Indian lineages. Gene-culture co-evolution was evident from the lactase-persistence allele being higher in pastoralists but lower in gold embossers despite shared ancestry. Noteworthy, despite the desert environment, the populations retain a high frequency of the SLC24A5 allele associated with lighter skin pigmentation in Europeans. Demographic analyses indicate an admixture of 60-80 generations before present and strong founder effects in certain groups, particularly tie and dye and Persian migrant artisans ∼500-600 years ago. Ancient DNA (aDNA) comparisons confirmed continuity with the Indus Periphery and historical South Asian populations. The genetic landscape of the Thar is a palimpsest shaped by successive layers of settlement, migration, and cultural continuity. By establishing the genomic baseline of Thar's craft and pastoral communities, this study shows how ecology and endogamy, along with population history, shape distinct genetic landscapes. These findings provide essential context for studying genetic risk, adaptation, and human resilience in extreme environments.

Attendance control and communication between parents and children are the most significant issues of contemporary learning settings. With the advent of a digital age, this can be robotized to cut a lot of administrative work and improve efficiency of work. The paper outlines a technological solution that will be a support of the attendance monitoring process in the classroom and real time exchange of data with parents. This system is proposed on Radio Frequency Identification (RFID) which is utilized in tracking the attendance of students. A special RFID tag has been provided to each student and scanned when he / she enters the classroom. An Arduino controller receives the data received and stores it safely and securely thanks to an SD card module. And it also comes with a built-in GSM system which transmits instant messages to parents to inform them about whether the child has gone or not. The system was constructed and was found to be functioning and has been quite efficient in capturing the attendance and timely communication facilities. The solution will reduce paperwork and manual record keeping which will assist in organizing the classroom and increase the clarity between the learning institutions and their parents. The paper will give detailed description on how the system was designed, how it was implemented and also an analysis of the system performance. The findings indicate that it can be used to improve parent-teacher relationship, reduce workload on the administration and overall improve attendance control. Besides, the study mentions potential improvements and increased use to introduce the system as a viable and realistic tool that could be utilised by any institution that needed to revamp their attendance system and communication.

The complex pattern of non-linear time series data is still difficult to predict with certainty even after series of developments in last centenary from statistical theories of 1930s to current trend of multivariate & spatio-temporal modelling that phased through non-linear models, machine learning and deep learning theories. In this research, prediction of Nifty 50 Index, a time-series data was undertaken by fusing Chaos theory with LSTM. The research has used chaos theory to reconstruct Nifty’s phase space by using embedding dimension and time delay to capture system dynamics, prepared training data to feeds it into an LSTM model to predict future values. Chaos theory is used to cleans and organizes the historical data and the LSTM learns how that history predicts the future. Chaos theory gives structured input features (capturing long-term dependencies in a deterministic chaotic system). LSTM learns the mapping from past states to future states. The system was used to predict the values of Nifty Index for very short-term period. An initial evaluation suggested the hybrid model produced lower forecasting error than Naïve, ARIMA, and standalone LSTM benchmarks. A subsequent closed-loop robustness check, prompted by a review of the chaos-feature construction, found that this advantage depended on whether the chaos-derived features were computed causally: under a strictly causal, look-ahead-free specification, the advantage did not replicate against a matched LSTM baseline, while a non-causal (look-ahead) construction reproduced an advantage resembling the original result. The principal contribution of this study is accordingly methodological: a fully specified, causal rolling-window procedure for embedding chaos-theoretic diagnostics into deep learning forecasting pipelines, and a demonstration of how readily an under-specified construction can manufacture an illusory performance advantage in hybrid chaos–deep-learning models.

This paper develops a two-period dynastic overlapping-generations (OLG) model in which parents simultaneously choose consumption, savings, fertility, and three distinct dimensions of child quality-education, physical health, and mental health-under a pay-as-you-go (PAYG) pension system. The central innovation is modelling mental health as an independent productivity-enhancing input with its own elasticity θ in a Cobb-Douglas human-capital technology. This yields simple proportional allocation rules and shows how pension policy affects not only the overall level but also the composition of human capital investments. In steady state, higher PAYG contribution rates raise fertility through the Yakita effect but crowd out per-child investments in all quality dimensions, including mental health. An increase in the mental-health elasticity θ shifts resources toward non-cognitive skill development while reducing fertility. These results reveal a fundamental policy tension for developing economies: pension systems that rely on children for old-age support simultaneously increase birth rates while reducing long-term human capital formation, with disproportionate effects on non-cognitive skills. The framework provides theoretical guidance for complementary policies that protect mental-health investments, with particular relevance for countries such as India where children remain a primary source of retirement security and mental-health services are underfunded.